Question

I have 5 numpy arrays within a single nxm array, with each array containing a set of random values. What I need is to print them according to their decreasing cumulative sums. For instance, my nXm array looks like:

 v = [[1,2,3,4], [2,3,4,5], [11,21,3,4], [4,33,21,1], [2,4,6,5]]

and what I need it to be ordered like is: the one having the highest cumulative sum comes first. I tried printing them according to argmax but it sneaks into all the elements of individual arrays and sorts them in a descending order..

Is there a way?

Answer 1

Assuming that by cumulative sum you mean total (there’s a cumulative sum function which returns something else), then you can do this both using the standard sort:

>>> v = [[1,2,3,4], [2,3,4,5], [11,21,3,4], [4,33,21,1], [2,4,6,5]]
>>> sorted(v, key=sum, reverse=True)
[[4, 33, 21, 1], [11, 21, 3, 4], [2, 4, 6, 5], [2, 3, 4, 5], [1, 2, 3, 4]]

and in numpy using argsort:

>>> a = np.array(v)
>>> a.sum(axis=1)
array([10, 14, 39, 59, 17])
>>> a.sum(axis=1).argsort()
array([0, 1, 4, 2, 3])
>>> a[a.sum(axis=1).argsort()[::-1]]
array([[ 4, 33, 21,  1],
       [11, 21,  3,  4],
       [ 2,  4,  6,  5],
       [ 2,  3,  4,  5],
       [ 1,  2,  3,  4]])

But I may be misunderstanding you.